Evaluate Clustering Performance Without Ground Truth Labels

In the absence of ground truth labels, evaluating clustering performance is difficult.

Yet, there are a few performance metrics that can help.

Using them, you can compare multiple clustering results, say, those obtained with a different number of centroids.

This is especially useful for high-dimensional datasets, as visual evaluation is difficult.

Silhoutte Coefficient:

• for every point, find average distance to all other points within its cluster (A)

• for every point, find average distance to all points in the nearest cluster (B)

• score for a point is (B-A)/max(B, A)

• compute the average of all individual scores to get the overall clustering score

• computed on all samples, thus, it's computationally expensive

• a higher score indicates better and well-separated clusters.

I covered this here if you wish to understand Silhoutte Coefficient with diagrams: The Limitations Of Elbow Curve And What You Should Replace It With.

Calinski-Harabasz Index:

• A: sum of squared distance between all centroids and overall dataset center

• B: sum of squared distance between all points and their specific centroid

• metric is computed as A/B (with an additional scaling factor)

• relatively faster to compute

• it is sensitive to scale

• a higher score indicates well-separated clusters

Davies-Bouldin Index:

• measures the similarity between clusters

• thus, a lower score indicates dissimilarity and better clustering

Luckily, they are neatly integrated with sklearn too.

• Silhoutte Coefficient

• Calinski-Harabasz Index

• Davies-Bouldin Index

👉 Over to you: What are some other ways to evaluate clustering performance in such situations?

👉 Read what others are saying about this post on LinkedIn and Twitter.

👉 Tell the world what makes this newsletter special for you by leaving a review here :)

👉 If you liked this post, don’t forget to leave a like ❤️. It helps more people discover this newsletter on Substack and tells me that you appreciate reading these daily insights. The button is located towards the bottom of this email.

👉 If you love reading this newsletter, feel free to share it with friends!

👉 Sponsor the Daily Dose of Data Science Newsletter. More info here: Sponsorship details.

Find the code for my tips here: GitHub.

I like to explore, experiment and write about data science concepts and tools. You can read my articles on Medium. Also, you can connect with me on LinkedIn and Twitter.